Social network theory has the potential to improve our understanding and treatment of human health issues on multiple levels, but we currently lack the basic information on how the spatial and mathematical relations of networks relate to the content and quality of relationships and how such variation influences health outcomes. We propose to determine, using a nonhuman primate, how internal (e.g., personality and temperament, genetic predispositions) and external factors (e.g., environmental and social stressors) in multiple individuals interact to affect network structure and dynamics and how these, in turn, influence health outcomes in social communities. We believe a nonhuman primate model offers several advantages to the advancement of social network theory with regard to human health because monkeys provide a cognitive and social analog for humans, data can be collected by direct observation of multiple communities (providing statistical replication), and the genetic and social history of all individuals is fully known.)We have four specific aims: (1) advance theory and methodologies assessing the network dynamics and robustness at multiple levels pertinent to the health context, (2) characterize how internal and external factors acting on individuals collectively influence network structure and robustness, (3) quantify the influence of network structure and robustness on metrics of stress as health outcomes and (4) assess the effects of experimental perturbation of network composition on network structure and robustness and health outcomes. Three main categories of data will be collected: (1) behavioral observation of affiliative and aggressive interactions, (2) assessment of individual internal factors including biobehavioral assessment of personality/ temperament and genotyping of the 5-HTTPLR and MAO-A genes, and (3) behavioral, physical and physiological measurement of health outcomes, including Rhadinovirus shedding, C-reactive protein levels, attitude, hydration, body condition, and trauma. Each of eight social groups will be observed 78 weeks across two years. Observers will record affiliate, aggressive, and submissive interactions among individuals using an event sampling design. Personality/temperament will be assessed by rating each animal on a list of 50 personality traits. Several health outcomes will be measured daily and during routine roundups. Behavioral data will be used to construct various social networks whose structure, dynamics, and robustness will be measured, and subsequently analyzed with respect to internal factors, health outcomes, and behavioral measures using multi-level generalized linear models.)